Post Go back to editing

Not able to lock ADF4350

Hi

For past few days i m working on ADF4350 but not able to achieve lock. I have designed the PCB on FR4 (.8 mm) based on the circuit diagram given in Application Information heading on Pg25 of datasheet. Bottom ground has been provided and since i am using only 1 RF output, the other complimentary output has been terminated with 50 Ohm. Auxilary output is mute. I designed the loop filter values using ADIsim PLL. Required output frequency is 500MHz. I am using Ref Freq=20MHz. Its square wave with 3.3V Amplitude generated by Arbitary function generator with High z output. Loop bandwidth is 100 KHz and PFD freq is 1MHz. Loop filter values comes as: C1 =22.6pF, R1=14.5K,C2=308pF,R2=29.6K,C3=10.3pF. In layout all precautions has been taken like placing decoupling close to supply line etc. I m transfering the control words using serial port of PC. I am using option of Digital Lock Detect, disabled mute till lock detect, prescalar is 8/9 , refrence divider and doubler both have been disabled and since my PFD frequency is 1MHz (>125KHz), i have selected Band Select Clock Divider to 16. I m trying to make ADF4350 work as Integer N PLL, so FRAC is 0. Furthermore Option for Low spur has been enabled.Since my required output should be 500MHz hence output divider is set to 8 and 4GHz of frequency is generated. Loop filter is passive and hence Phase detector polarity is set to 1. Rset is 4.7K and hence Icp is 5 mA. My counter values are as follows:

Output Freq Req: 500MHz    Ref Freq: 20 MHz,     PFD Freq: 1MHz, Output Divider of 8 Modulus selected

R =20 and N=4000

I m sending data and clock thru PC serial port by programing it in C. Programing has been done to take care of Data to CLK setup time Data to Clock hold time CLK to LE setup time, LE pulse width etc. I have verified these waveforms on Oscilloscope and its correct. Initialization sequence is Register 5 -4-3-21-0. Data is transfered as MSB first. Following is my Initialization sequence:

                 MSB                                                                                LSB

Register5:    0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1

Register4:    0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 1 1 1 1 0 0

Register3:    0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1

Register2:    0 1 1 1 1 0 0 0 0 0 0 0 0 1 0 1 0 0 0 1 1 1 1 1 1 1 0 0 0 0 1 0

Register1:    0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1

Register0:    0 0 0 0 0 1 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

With this control words i m nt getting lock. I m seeing various spurious signals on spectrum analyser and there is a signal of 532 MHz , 0dBm but it doent seems to be phase locked as in low span there is a considerable amount of jitter.

 

Please help me troubleshoot this problem. I have worked on LMX2326 (National Semiconductor) but never faced such kind of problem.  Please help meo out.

 

Regards

Piyush

  • Please remove the 50 Ohm resistors from the RFOutB+/- pins, as these should be left open if unused.

  • Also, do not use the divide by N on MUXOUT as it affects the band select logic.

    If you absolutely must use it, then program R0 to update the frequency first. Then write to R2 to program MUXOUT to N div mode. This way band select is unaffected by the MUXOUT state.

  • Can you connect an oscilloscope to the muxout pin, and then program the various muxout options from Figure 26 on the datasheet? This will verify if you are actually writing to the part.

  • Hi rbrennan

    Thanks for the reply. I forgot to mention this point earlier. I tried with various mux out option viz R divider, N divider etc. I got 1 MHz output for R divider selection which is correct. But for N divider i was getting several components on oscilloscope. may be because the PLL is not locked. what should i do?.. i rechecked my control words and that seems fine. what can be the possible error?
    Thanks again for ur reply and hoping to get a clue to solve this problem
    bye
  • Hi Ian

    I have removed this 50 ohm resistor, but then also the output is same. what can be the other possible mistakes which i am making. Is my control words correct?. i have rechecked it several times and it seems right. I am getting divide by R output from Mux pin. But i am not able to get N divider output. I forgot to mention this in my earlier post. Can u please help.

    Waiting for a reply

    thanks

    bye

  • Hello Piyush,

    Can you check the contents of the registers, in particular R4, in your initial email , you wrote 0080143C for R4...this gives an o/p divide of 1 instead of required 8.

    You can download the ADF4350 eval software from our ADF4350 product page, and plug in your register values, this will give you a good indication that your calculated register values are correct.

    Regards,

    Brigid.

  • Hello Piyush,

    As you mention that you are not getting the N divider o/p, can you  check the voltage at the Vtune pin.

    regards,

    Brigid.

  • Greetings,

    I recently spent several hours trying to get an ADF4351 to work properly while showing the N divider at the mux output, in an attempt to track down the source of an occasional phase-jump problem.  I now see what I need to do, based on the band select problem on the ADF4350.  The information is is very valuable, but sometimes difficult to find. 

    I have not been able to find an errata list for the ADF4351, or for the ADF4350, for that matter.  Is this site the best (and only) source of this kind of information? 

    Thanks for your help,

    lig

  • Hi,

    We generally do not recommend that MUXOUT = Ndiv out be used for anything but debug, so I guess it had not occured to us to document this. I will however put a sentence about this on the ADF4351 datasheet.

  • This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin